{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-99621","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-09-25T10:25:14.425Z","datePublished":"2026-11-20T05:49:41.407Z","dateUpdated":"2026-11-20T05:49:41.407Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-11-20T05:49:41.407Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu: unlock the VM and BO when attaching a KFD BO fails\n\nkfd_mem_attach() locks the page directory and the BO with a drm_exec\ntransaction on its stack, but leaves through the unwind label without\ndrm_exec_fini() when locking fails, e.g. when MAP_MEMORY_TO_GPU gets a\nsignal while the page directory or the BO is busy. Reservations already\nacquired then remain locked with their ww_acquire_ctx pointing into the\nexpired stack frame.\n\nWith one thread doing GEM_VA map/unmap on the KFD VM and\nanother one mapping KFD BOs while getting SIGUSR1 every 20 us:\n\n  BUG: KASAN: stack-out-of-bounds in __ww_mutex_lock.constprop.0+0x34f5/0x3da0\n   ww_mutex_lock_interruptible+0x2f/0x150\n   drm_exec_lock_obj+0x107/0x2d0 [drm_exec]\n   kfd_mem_attach+0x4dd/0x1b60 [amdgpu]\n\nand the process ends up in D state.\n\nFinish the transaction before unwinding.\n\n(cherry picked from commit 7b885fa27e4b9fa5cbdc64e66fc35872aee296f5)"}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c"],"versions":[{"version":"fd1fa48b935f22c7d99713bf33846e14a6bb6ab9","lessThan":"1ae594845217ce740b27ec5a202e9e8cb144e0aa","status":"affected","versionType":"git"},{"version":"fd1fa48b935f22c7d99713bf33846e14a6bb6ab9","lessThan":"e0a1404b34e0002f03f7979cdb791a9b5c3963f5","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c"],"versions":[{"version":"7.0","status":"affected"},{"version":"0","lessThan":"7.0","status":"unaffected","versionType":"semver"},{"version":"7.3.5","lessThanOrEqual":"7.3.*","status":"unaffected","versionType":"semver"},{"version":"7.4-rc2","lessThanOrEqual":"*","status":"unaffected","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"operator":"OR","negate":false,"cpeMatch":[{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"7.0","versionEndExcluding":"7.3.5"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"7.0","versionEndExcluding":"7.4-rc2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/1ae594845217ce740b27ec5a202e9e8cb144e0aa"},{"url":"https://git.kernel.org/stable/c/e0a1404b34e0002f03f7979cdb791a9b5c3963f5"}],"title":"drm/amdgpu: unlock the VM and BO when attaching a KFD BO fails","x_generator":{"engine":"bippy-1.2.0"}}}}